AMPK: The Perfect Partitioning Agent?

buff guy bar curling
AMPK: The Perfect Partitioning Agent?
by: Wen Zhang

Recap from last month

termining what functions they were optimized for. In assuming that a metabolically healthy organism is one whose tissues are processing fuels (in a manner consistent with both its specified functions and environmental signals), a point of reference for ‘good’ partitioning could be established. All manipulations that altered this state therefore had to be qualitatively evaluated by looking at the stressor (environment) and the metabolic flexibility of the organism. Thus, what might be constituted as beneficial from a tissue-function point of view could be detrimental in a certain environment. For example, skeletal muscle is ideally suited for lipid oxidation and glycogen storage. If we administered a mixed macronutrient meal to an animal, we would consider it normal if its skeletal muscle took up the majority of the glucose for storage as glycogen and oxidized the lipids for fuel. If we gave the animal an excess of fuel (and the muscle accordingly increased its glucose storage and lipid oxidation to quickly clear nutrients from the bloodstream), we might even classify the actions as positive partitioning. And we would be correct if the animal had been consuming adequate quantities of food for some time. However, in a starving organism (where glucose is lacking and fuels are depleted), glycogen is the last place we would want glucose to go, as there are other less metabolically flexible tissues which demand glucose for survival (i.e. CNS tissue). So, when discussing good vs. bad and positive vs. negative partitioning, we must first define the terms we apply – which in this case is the overall metabolic flexibility of the organism and the environment to which it is being exposed.


Bodybuilding and the metabolic syndrome share several similarities, yet the end results are, to put it mildly, different. MSX can be incited through a chronic excess influx of nutrients; building muscle similarly requires fuel above and beyond what is needed to maintain the status quo. MSX is tightly linked to inflammatory responses, while the training necessary to stimulate muscle growth is known to generate massive quantities of free radicals and inflammatory cytokines [1]. Insulin resistance is a classical symptom of MSX, and heavy resistance training is known to induce temporary insulin resistance [1]. These parallels guide us to a central question, and the topic of this month’s article: why do two metabolic adaptations that share similar environmental factors lead to such different phenotypes?

At this point, most readers are probably thinking “it’s exercise and diet, genius- bodybuilders eat clean foods and train, fatties with MSX don’t.” And this would be essentially correct. However, once we go beyond the obvious, we are still left with questions concerning the underlying mechanisms. After all, what body composition enthusiast has not wondered about which exercise regimen or diet plan will yield optimal results (optimal results meaning a higher muscle:fat ratio)? To understand how we can optimize our approach, we must first understand why the body reacts as it does. Towards this end, there are two primary variables to consider – diet and exercise. We will focus on the latter for now.

From the viewpoint of a cell, exercise is a stressor that demands certain adaptations be made. The need for adequate fuel supply to support contraction, establishment of efficient motor unit recruitment and control, and upregulation of free-radical quenching systems are all examples of exercise-induced adaptations. It is a well-known fact that the extent of these changes is directly related to the magnitude of the demand placed upon the body, up to a certain physiological threshold. This implies that sensors are present within cells, which translate the mechanical and chemical stresses of exercise into meaningful adaptations, presumably in a dose-dependent manner. Considering that the metabolic alterations induced by exercise lead to increased metabolic flexibility and a generally improved ability to cope with environmental stressors, there is much to be gained if we can discover the identity of the sensors, how they are activated, and what we can do to enhance or better control their regulation.

One of the most serious issues the cells of an exercising organism must deal with is maintenance of fuel supplies. The animal must be able to store enough fuel to keep up with the increased energetic demands and utilize the fuel rapidly. Intuitively, it makes sense that the severity of the fuel depletion per unit time must have some influence on the type and magnitude of adaptations a cell would be forced to make. After all, a cell would not want to under-compensate any more than it would wish to over-compensate. Additionally, it would also be expected that cell types most directly affected by the stresses of exercise would be most likely to adapt. We must then ask, from the perspective of fuel partitioning under caloric excess plus exercise, which fuel depletion approach would be most efficient, and which tissues we should focus our attention on.

To briefly summarize, we are now faced with the following scenario: we have an organism in its basal state – that is, it is healthy, and in energetic equilibrium with its environment. We now introduce two stressors – excess fuel and exercise. The questions we need to answer are:

1. Do different types of exercise lead to more effective handling of metabolic fuels?
2. What tissues are most affected by exercise?
3. What sensors in the cell mediate exercise-induced adaptations?

If we were trying to find solutions to the questions above in the absence of any prior knowledge, we would do best by starting from #1 and working in order to #3. Fortunately, much work has already been done in this field, so we can change our approach. The third question is potentially the most significant, because answering it will reveal to us the rationale behind adaptations as well as how to manipulate these changes at the molecular level.

As stated above, the most serious problem presented to exercising cells is keeping an adequate flow of energy to support contraction at a specified intensity. To control this, cells have evolved an elegant sensing mechanism that is responsive to several signals, all of which are related to changes in the energy status of the cell and/or to exercise-associated signals. And, because they are directly involved in the contraction process, skeletal muscle cells will be the focus of our analysis.

[Note: The late Karl Hoffman wrote an excellent article focusing on AMPK in a previous issue, which readers are encouraged to revisit.]


Introduction to the adenosine-monophosphate-activated protein kinase, AMPK

The enzyme that was later named AMP-kinase was first observed in 1973 by Beg et al. as a protein which inactivated the HMG-CoA reductase, a key enzyme in sterol synthesis; in the same year, Carlson and Kim observed inactivation of liver acetyl-CoA carboxylase by an “ATP-dependent kinase”[1]. Over a decade later, Carling et al. made the connection between these two phenomena, leading to the conclusion that one kinase, responsive to AMP, mediated this effect [2]. Subsequently, this kinase was named “the AMP-activated protein kinase”, or AMPK [3]. Since then, AMPK has been termed as a “metabolic fuel sensor” [4] as well as a “fuel gauge” [5]. When one looks at the numerous enzymes that AMPK controls and the conditions under which it is most active, it becomes apparent that this title is well deserved.

Having a n enzyme that senses changes in AMP status as opposed to ATP allows much tighter regulation of cellular energy charge [6]. T he adenylate kinase (AK) reaction (2ADP ß à ATP + AMP) allows cells to maintain a 1:10 ADP: ATP ratio; by mass effect, the AK reaction favors the reverse direction, generating ADP and preventing a high concentration of AMP from accumulating. In this manner, the AMP:ATP ratio is maintained at approximately 1:100 [6]. However, when ATP consumption exceeds ATP synthesis, as occurs in hypoxia, ischemia, exercise, and with uncoupling of oxidative phosphorylation from ATP synthesis, the ADP: ATP ratio decreases, which forces the equilibrium of the AK reaction in the forward direction, generating ATP and AMP [6]. Because the concentration of ATP is normally much higher than ADP or AMP, the AMP:ATP ratio is low (1:100) at rest. This means that when the AK reaction favors the forward direction, the relative increase of AMP (as a percentage of original) will be much greater than that of ATP, since the adenylate kinase generates equimolar amounts of ATP and AMP from 2ADP molecules. This means that a 10-fold increase in the ADP: ATP ratio results in a 100-fold increase in the AMP: ATP ratio, a difference of one magnitude [6].

Physiological effects – overview

It is necessary to know the details of AMPK sub-unit regulation, as our goal is to optimize nutrient partitioning by better understanding how exercise works its magic, and AMPK is obviously a critical exercise-sensitive signal. However, before we explore further how exercise-induced signals affect AMPK, we should first look at the actual effects of AMPK on nutrient partitioning to answer an ever-important question: “so what?”

AMPK exerts immediate and long-term effects through regulation of enzyme activity and expression, meaning that it acts quickly through phosphorylation of substrates, and more slowly (but also more enduringly) via nuclear mechanisms at the level of the genome. Enzymes regulated by AMPK include the HMG-CoA reductase, acetyl-CoA-carboxylase, glycerol-phosphate-acyl transferase (GPAT), endothelial nitric oxide synthase (NOS), 6-phosphofructo-2-kinase, malonyl-CoA-decarboxylase, stearoyl-CoA-desaturase, and hormone-sensitive-lipase [6]. This implicates AMPK in the regulation of free fatty acid, triacylglycerol, phospholipid, sterol, and nitric oxide synthesis, as well as regulation of glycolysis and gluconeogenesis. These effects may be direct, mediated through other proteins/transcription factors (SREBP-1, PPAR-gamma, PGC-1a), or a combination of both. In the long term, fatty acid synthesis and TAG synthesis are both repressed through reduction of ACC, FAS, SCD-1 and S14 expression/MCD upregulation. Gluconeogenesis is dampened through reduction of L-type-pyruvate kinase and PEP-CK expression and glycolysis is upregulated due to increases in expression and activity of hexokinase, GLUT4, and a number of mitochondrial enzymes involved in oxidative phosphorylation [6, 26, 27]. The expression of UCP-3 is also increased with chronic AMPK stimulation, possibly protective effect against increased beta-oxidation due to AMPK, and AMPK’s potentiation of PGC-1a may promote the activity of several free-radical quenching enzymes [27].

The fact that AMPK is activated by a change in energy status, and has a net effect of increasing beta-oxidation of fatty acids, reducing synthesis of lipids, improving glucose clearance, inhibiting hepatic glucose output, upregulating mitochondrial respiratory chain enzymes, and enhancing cytoprotective proteins; suggests that this enzyme meets all of the requirements of an ideal cure of the problem of excess fuel influx as discussed in the previous installment. Like most enzymes, though, AMPK is not a single-isoform protein; in fact, it is a three-component modular protein consisting of different combinations of three subunits, each with two or more isoforms. We will first examine how sub-units interact with each other, the unique characteristics of isoforms, and whether there are certain combinations that are positively associated with the physiological effects we seek.

Locations of AMPK isoforms and heterotrimer combinations

AMPK is a heterotrimeric serine/threonine protein kinase composed of an alpha, beta, and gamma sub-unit and is ubiquitously expressed in all mammalian cells [7]. AMPK controls, in the short term through phosphorylation, enzymes such as ACC, HMG-CoA reductase, glycogen synthase, HSL, and several others, as well as exerting effects on cell proliferation and differentiation via nuclear effects [8. Activation of AMPK occurs in response to signals that are the result of impaired energy status, and co-expression of all three subunits is required for maximal catalytic activity [9].

In mammals, there exist two isoforms of AMPK-alpha (a1/a2), two of AMPK-beta (b1/b2), and three of AMPK-gamma (y1/y2/y3), allowing for at least 12 combinations of the AMPK heterotrimer; each subunit is coded on a separate gene [6].

The alpha subunit of AMPK is the catalytic portion of the heterotrimer. Both isoforms of the alpha-subunit are 63-kDa proteins consisting of a kinase domain at one end, and a regulatory region at the other. The regulatory region contains an autoinhibitory section that prevents kinase activity unless AMP is present [10]. Based on yeast studies, the regulatory region is believed to bind to the gamma-subunit, and the gamma–subunit in turn binds to the beta-subunit [6, 8]. Both alpha1 and alpha2 bind beta and gamma subunits in a 1:1:1 ratio [9]. AMPK-alpha1 is ubiquitously expressed, while, alpha2 is expressed mainly in the skeletal muscle and cardiac tissue, as well as in liver. The alpha1 and alpha2 subunits have 90% homology within their catalytic cores, and have 64% homology to the yeast equivalent, Snf-1p, showing that AMPK is an evolutionarily conserved enzyme [8].

The beta-subunit is a smaller (38kDa) protein [4] and contains three regions- the N-terminus, the KIS (middle), and the C-terminal ASC domains [6]. The alpha-subunit’s catalytic domain binds to the middle segment, while the gamma-subunit binds to the ASC section; the beta-subunit thus mediates localization of the alpha and gamma subunits, and may facilitate AMP binding at the interface of the alpha and gamma subunits

The beta1 isoform is widely distributed, while beta2 is seen mainly in muscle [11, 12]. The beta-subunit is myristoylated and phosphorylated at multiple sites, and although the exact function of this is not yet known, myristoylation generally targets proteins to hydrophobic regions of the cell, i.e. membranes (although a mutation of the myristoylation point does not fully eliminate membrane binding of the AMPK [8]). Phosphorylation at up to four sites may additionally control subcellular localization of the beta-subunit [13].


The gamma subunit is a similar in size to the beta-subunit (35kDa) [4] and has regulatory functions by forming an AMP-binding pocket with the alpha subunit [6]. While hypothetical, evidence that the AMP-binding site is not located on the alpha- but rather on the gamma-subunit, exists [13]l in any case, the gamma subunit is essential for optimal alpha-subunit catalysis. AMPK-gamma1 and gamma2 are widely expressed, while gamma3 is exclusively found (thus far) in skeletal muscle, although only as a minor constituent depending on the organism in question [14]. The gamma-subunit is modified only at the N-terminal by acetylation, and the entire protein consists mainly of 4 sequence motifs termed the cystathionine beta-synthase domains (CBS), the functions of which have not been elucidated [8]. A paper published by Adams et al. in 2004 presented evidence that the CBS domains of the gamma subunit are needed to render the alpha subunits inactive in the presence of ATP [28]. Discovery of interaction of the CBS domain with S-adenosyl-methionine suggests that the gamma-subunit’s four repeating CBS domains may be allosterically modified by S-adenosyl-methionine, a finding whose significance remains to be established [6].

Due to the many potential combinations of subunits, AMPK might be differentially regulated in a tissue-specific manner based upon different subunit configurations. For example, l iver primarily consists of alpha1-beta1-gamma1 as well as alpha2-beta1-gamma1 combinations, while the abundance of the beta1 subunit is low or nonexistent in skeletal muscle; conversely, the beta2 subunit may be exclusive to skeletal muscle [12]. Why skeletal muscle lacks detectable beta1-containing AMPK heterotrimers in contrast to liver is unknown, but if this trend holds true across tissues, it makes possible the notion of tissue-specific AMPK activators.

Sub-unit specificity

The unique distribution, combination, and regulation of AMPK sub-units has prompted extensive research into the question of whether physiological effects could be linked to specific AMPK subunits. Human data is somewhat conflicting on this issue, possibly due to difference in methodology and subject recruitment. In type II diabetics subjected to six weeks of 30-minute lower body resistance training exercise 3x/week at 50% 1RM for the first three weeks and 70-80% 1RM for weeks three to six, muscle biopsies from the vastus lateralis (taken 15 hours after the last exercise bout) showed significant increases in protein content of the a1 (16%) and y1 (29%) sub-units only, while the level of the y3 subunit decreased (48%). Interestingly, the protein content of AMPK sub-units were not different between controls and T2D subjects before or after resistance training, and the magnitude of changes in protein expression were identical, implying that T2D does not affect actual protein content of AMPK sub-units in skeletal muscle [15]. This study, however, used older subjects (average age = 60) and a rather mild training protocol – therefore findings may be different in younger and/or trained humans. Additionally, the authors did not look at AMPK activity, which arguably is the main issue of importance.

The authors of the above study concluded that in human vastus lateralis muscle, the predominant AMPK heterotrimer combination is alpha2-beta2-gamma1/3. While alpha1, beta1, gamma2a, and gamma2b are all expressed in this tissue, the authors were unable to find beta1, gamma2a, or gamma2b in any complexes with alpha1 or alpha2 [15]. The significance of these selective changes are currently under investigation by several laboratories.

When young males were subjected to a single twenty-minute bout of cycling at 80% of VO2 max, Nielsen et al. found that the activity of the a1 subunit was not increased in response to exercise, while the a2 activity did increase- this effect was seen in both trained and untrained subjects [16]. This could be due to the fact that trained subjects were much better able to maintain their phosphocreatine:creatine ratio, and the PCr:Cr balance has been shown to affect AMPK activity, with a lower ratio (indicative of a lower energy state) being inversely proportional to AMPK activity [19]. Consistent with the previous study, though, the y3 sub-unit’s protein expression tended to be lower in the trained group; the a1 sub-unit protein levels were also significantly elevated in trained vs. untrained. Since this study used only a single exercise session, it supports other findings that expression of the a1 subunit is increased[15-17, 29], and the y3 subunit protein expression decreased in trained vs. untrained humans [15-17]. Again, the actual significance of this phenomenon is currently unknown.

A similar study using young, untrained males subjected to three weeks of single-leg extensor exercises at 70-85% of peak workout load (15 sessions over 3 weeks at 1-2 hours per session), found that protein levels of the a1, b2, and y3 subunits increased significantly, while that of the y3 decreased [17]. Additionally, activities of both the a1 and a2 subunits increased (94% and 49%, respectively). What is most interesting about this study, though, is that muscle biopsies were taken at 15-hours and 55-hours after the last exercise session at the end of the study. Therefore, the authors demonstrated that both AMPK protein levels and actual activity of the catalytic subunit persist far after the conclusion of exercise, with chronic training [17]. The observation of increased activity in both the a1 and a2 subunit are inconsistent with Nielsen et al., but are supported by an experiment conducted by Clark et al. where highly trained endurance athletes were subjected to three weeks of high-intensity training [18]. Over the course of three weeks, athletes in this study completed seven HIT trials consisting of cycling at 85% of VO2max, with each session involving eight five-minute intervals. At the end of three weeks, there was a decrease in CHO oxidation and a concurrent increase in fat oxidation during a 90-minute endurance trail compared to pre-HIT, and both a1 and a2 AMPK were activated, although there was no observable effect of this parameter due to HIT [18]. Again, it is possible that this was due to better maintenance of acute energy stores, as muscle lactate decreased by almost two-fold due to HIT. Note that these effects were all seen in highly trained endurance athletes, suggesting that high-intensity exercise offers metabolic benefits above and beyond that attainable with lower-intensity protocols. It is difficult to determine if AMPK was responsible for these effects, as no direct evidence was presented that correlated the changes with HIT-induced increases in AMPK activity; again, we can only try to rationalize this lack by assuming that any increases in AMPK activity were masked by the improved maintenance of phosphagen stores, as indicated by reduced muscle lactate values. Had biopsies been performed at the end of each week, we would be able to better tease apart AMPK’s contribution(s).

In conclusion, human studies show that the predominant AMPK heterotrimer combination in skeletal muscle is a2-b2-y1. The protein content of individual subunits changes in response to exercise, with long-term training inducing the expression of a1 and possibly y1, while decreasing the expression of y3. The activity of AMPK is affected by training intensity, as protocols that place a heavy energetic demand on cells seem to be the only instances where a1 activity is significantly upregulated [17-20], while a2 AMPK seems to be activated even at lower intensities [16, 20, 21].

Targets of AMPK

Coming back to the physiological effects of AMPK activation, a list and brief description of some of the more well-studied targets of AMPK is presented.

Acetyl-CoA carboxylase

Acetyl-CoA carboxylase (ACC) is a cytosolic enzyme that exists in two iterations- alpha and beta, or 1 and 2. The two isoforms have different sensitivities to substrates and intracellular localization, and thus it is important to distinguish which isoform is under scrutiny when reading studies.

The function of ACC is to generate malonyl-CoA from acetyl-CoA in the cytosol during the process of de novo lipogenesis. Even though DNL is of minor consequence in terms of the absolute quantity of fatty acids synthesized, it may have important implications through localized metabolite formation. Additionally, the flux of acetyl-CoA through ACC is a potent inhibitory mechanism with respect to beta oxidation; this is one manner in which the body decreases its rate of fatty acid oxidation when carbohydrates are readily available, as cytosolic acetyl-CoA relies upon a steady efflux of citrate from the mitochondria to the cytosol, and citrate synthesis is largely dependent upon glycolysis. Thus, high CHO oxidation à increased mitochondrial citrate à increased cytosolic citrate à increased cytosolic acetyl-CoA à increased malonyl-CoA à inhibit CPT-I à inhibit passage of LCFA into mitochondrial matrix à inhibit beta-oxidation. That AMPK can phosphorylate and inactivate ACC, and particularly the beta-or-“2” isoform, which is localized to the mitochondria, implies that it is able to relieve the inhibition of beta oxidation normally seen when CHO oxidation is elevated.


ACC is phosphorylated by AMPK at three residues: Ser79, 1200, and 1215; Ser79 seems to be the site of primary importance in regulating ACC., while Ser 1200 and 1215 may be less important, evidenced by the finding that a mutant ACC expressing Ala79 is resistant to inactivation by AMPK [4]. In hepatocytes, activation of Ser79 of AMPK via heat shock or AICAR almost totally prevents fatty acid synthesis in hepatocytes; both ACC1 and ACC2 are inactivated by AMPK [4, 22].

AMPK phosphorylates Ser79, 1200, and 1215 in ACC-1, while PKA phosphorylates Ser77 and 1200 (8). Either kinase will increase the Ka of ACC-1 for citrate by 100%, but AMPK phosphorylation additionally leads to an 80-90% decrease in the Vmax for ACC-1; PKA causes a mere 15% decrease in ACC-1 Vmax [23]. Because removal of ACC-1 n-terminus fully reactivates ACC-1, Ser77 and Ser79 emerge as the more important phosphorylation sites [23]. Further observations that Ser77 is not phosphorylated in vivo have led to the conclusion that Ser79 is the primary regulatory site, which is controlled only by AMPK. This suggests that in vivo, AMPK is the primary regulator of ACC-1 (and ACC-2), while PKA may have a negligible role by comparison [23], and there is no evidence that PKA phosphorylates/controls ACC in vivo [22].

ACC-2 is phosphorylated by AMPK, with resulting increases in Ka for citrate and a decreased Vmax, as well, meaning that ACC-2 will require a higher concentration of citrate for activation, and has a lower maximal catalytic rate [23]. Because of ACC-2’s elongated N-terminus, the Ser79 equivalent is located on Ser218 in ACC-2 [23]. The alpha2 AMPK is the main form of the catalytic subunit in skeletal and cardiac muscle, which is also more dependent on AMP than AMPK-alpha1; this combined with increased sensitivity of ACC-2 to ATP is complementary to the energy sensing function (as opposed to storage) of these two enzymes in muscle cells [23]. Additionally, heart and skeletal muscle have an isoform of CPT-I that is more sensitive to inhibition by malonyl-CoA, implying stronger response to AMPK activation (alpha2 isoform).

AICAR in vivo decreases ACC-2 by 50%, and malonyl-CoA levels by 40% in isolates rat soleus muscle [23]. It also increased fatty acid oxidation by 90%, suggesting that small changes in malonyl-CoA result in larger changes in beta-oxidation in skeletal muscle, although application of this data to other cells must be made with the realization that muscle is especially sensitive to malonyl-CoA, at the level of ACC-2 and CPT-I isoforms. However, the finding that incubation of hepatocytes, which express ACC-1 and ACC-2, as well as AMKP-alpha1 and AMPK-alpha2 equally, with AICAR, leads to inactivation of ACC-1 and 2, is encouraging [23].

Hormone sensitive lipase

Hormone-sensitive lipase is an essential enzyme that mediates the lipolysis of long-chain fatty acids from their esterified form to their free-form, allowing intracellular oxidation. Traditionally, PKA through the cAMP cascade has been considered to be the primary activator of HSL activity. AMPK seems to have effects on HSL as well, although the result has been found to either be inhibitory or neutral.
HSL is phosphorylated at Ser563 by PKA, and Ser565 by AMPK; phosphorylation at Ser565 by AMPK prevents PKA phosphorylation at Ser563, which is necessary for activation, suggesting that AMPK exerts an anti-lipolytic effect by preventing HSL activation by PKA [4]. While this is counterintuitive at first, this may have the ATP-conserving function of limiting lipolysis to a rate that prevents futile cycling of FFA (or ketogenesis), to match the rate of beta-oxidation. It also may contribute to the reduced plasma lipid profiles seen under conditions in which AMPK is activated AICAR inhibits the beta-adrenergic stimulated lipolysis in adipocytes through antagonism of HSL, as beta-adrenergic activation activates HSL through cAMP-PKA [4].

HMG-CoA reductase

The HMG-CoA reductase is the rate-limiting enzyme in cholesterol biosynthesis, and as such, can be important in regulation of plasma lipids. It is also interesting to see how an essential process, such as sterol synthesis, is given higher energetic priority versus a non-essential one, such as lipid synthesis.

AMPK appears to phosphorylate and inactivate HMG-CoA reductase through phosphorylation of Ser871, as mutation of this residue to an alanine eliminates AMPK inactivation, although it still remains sensitive to end-product negative feedback, suggesting that AMPK regulates HMG-CoA reductase through a mechanism independent of substrate availability, in accordance with AMPK’s role in mediating nutrient/stress regulation [8].

Activation of AMPK by heat shock, arsenite, or AICAR in rat hepatocytes results in a dramatic decrease in sterol synthesis [4]. Deoxyglucose treatment also completely inhibits sterol synthesis with wild-type HMG-CoA reductase, Ala871 is resistant to deactivation by 2-DG è although a mutant expressing Ser871 [4]. Phosphorylation of HMG-CoA reductase by AMPK, however, does not completely eliminate its activity, as a mutant with sites optimal for AMPK phosphorylation is phosphorylated 5x more than wild-type, yet it retains some activity, showing that HMG-CoA reductase is not as sensitive to AMPK inactivation as ACC is, pointing towards the importance of maintaining sterol synthesis in times of stress, compared to the expendable nature of fatty acid synthesis [4].

AMPK may normally be attached to the cell membrane, and detachment allows it to phosphorylate and inactivate HMG-CoA reductase, which does not occur when AMPK is membrane-attached [24]. Membrane attachment is likely to be mediated through beta-subunit myristoylation [13].

C/EBP-beta/alpha; PPAR-gamma

AMPK-a2 can translocate to the nucleus, and we would therefore expect to see some effects at the nuclear level. While the following data is in vitro, it is intriguing that AMPK might inhibit preadipocyte maturation.

AICAR added at day 0 at 1mM to adipocytes prevents neutral TAG accumulation, although the effect is not seen at to a significant extent with only 0.25mM; the effect is diminished even if added at 1mM on day 3, and no effect is seen at all at 1mM if added at day 5 [11], pointing for exertion of control early in differentiation by AICAR, presumably through AMPK. The concentrations of AICAR used were chosen because maximal activation of AMPK is seen at [AICAR] between 0.5 and 1mM, while only minimal stimulation is seen with 0.25mM AICAR [11].

The earlier cells are treated with AICAR at sufficient concentrations (>0.25mM), the more dramatic the inhibition of lipid accumulation [11]. The reduction in lipid accumulation seems to be a byproduct of decreases in FAS and ACC, both of which are significantly reduced in expression following early treatment with AICAR, although the inhibition of ACC is somewhat more pronounced than that of FAS [11].

In addition, there are several transcription factors that are implicated in fibroblast differentiation into adipocytes, and these include members of the CAAT/enhancer binding protein (C/EBP) and the PPAR families (10). Studies have revealed that C/EBPbeta, followed by C/EBPgamma, PPAR-gamma, and finally, C/EBPalpha are expressed as the fibroblast differentiates [11]. PPAR-gamma is expressed by day 3, and reaches a peak by day 5-6. C/EBPalpha is seen by day 5-6 as well, and both PPAR-gamma and C/EBPalpha are reduced in expression by AICAR at 1mM [11]. In addition, AICAR reduces cell count by 30% by the third day, with 1mM addition [11].


While AICAR actually prevents the normal decline in C/EBPbeta that normally occurs by the second day, C/EBPalpha expression is reduced with AICAR, leading to the conclusion that AICAR is somehow causing an inactive C/EBPbeta to be transcribed, which does not allow subsequent induction of C/EBPalpha [11]. Thus, AICAR not only prevents ACC and FAS expression, but also interferes with cell differentiation factors C/EBPbeta and PPARgamma, and although the mechanism of interference has yet to be elucidated, AICAR’s effects in other studies strongly suggest that it is working through AMPK potentiation, although exactly how remains a question [11].

Sterol response element protein-1C

SREBP-1c has been referred to as a master lipid regulatory hormone. While this title is perhaps a bit much in light of new discoveries surrounding the PPARs and PGC-1a, it is nonetheless an important mediator in the lipogenic effects of insulin, and its over-activation is closely linked to obesity and MSX.

AMPK activation suppresses SREBP-1c and insulin-stimulated transcription factors that act through SREBP-1c [25]. Metformin prevents the induction of mature SREBP-1 accumulation in response to refeeding, as does AICAR [25], which suggests that AMPK regulates SREBP-1 to a significant extent. The role of SREBP-1 in lipogenic gene expression is indisputable, and there is much overlap between SREBP-1 induced genes, glucose-induced genes, and other enzymes that AMPK directly controls the activity and expression of, making SREBP-1 a potentially potent mediator of AMPK’s effects.


PGC-1a is a nuclear transcription factor coactivator, meaning that it aids transcription factors in mediating gene transcription. PGC-1a has been called a master metabolic regulator due to its profound effects on overall energy metabolism in the cell, and it appears that AMPK activates this transcription factor [27]. In this manner, AMPK ensures that downstream cellular metabolism will be sufficient to properly process the increased flux of high-energy intermediates of enhanced lipid metabolism. We will go into more detail about PGC-1a in the future, and indeed, an entire article will be dedicated to its functions.

Peroxisome-proliferator-activated receptor-alpha

Long-time residents of the Avant Labs forums are familiar with PPARs, a family of transcription factors that are the ultimate mediators of gene effects stimulated by upstream signals. PPAR-a is a crucial transcription factor for a number of gene involved in lipid metabolism, both mitochondrial and peroxisomal. It is unclear whether AMPK works through PPAR-a solely, or if activates parallel pathways. Adding AICAR to cells clearly increases fatty acid oxidation and the mRNA of PPAR-a; the increase in fatty acid oxidation is abolished in PPAR-a knockout mice and those treated with siRNA against PPARa, effectively silencing the gene [26]. We will discuss PPAR-a in conjunction with PGC-1a as they relate to AMPK in future articles.


Looking at all the targets listed above (the list is by no means exhaustive), several common themes emerge. First, we see that AMPK increases the ability of skeletal muscle to oxidize fats. Striated muscle is already very well-suited towards utilization of this fuel, and exercise seems to increase muscle’s threshold even more. One could argue that this effect is driven by a desire to conserve glucose, and in endurance-trained athletes who are introduced to a high-intensity exercise protocol for three weeks, this is indeed the case [18]. However, this adaptation serves two purposes – it potentially spares glucose during exercise, but it also upregulates beta-oxidative machinery, even days after the last exercise session with chronic training [17]. The interaction of AMPK with PGC-1a suggests that AMPK is able to increase not only genes related to fat oxidation [26, 27], but all the way down the respiratory chain, which may explain the ability of trained athletes to better maintain their high-energy phosphagens under intense exercise. Furthermore, AMPK increases glucose uptake independent of insulin. Even though it may inhibit glycogen storage, the net effect seems to be an enhancement of glycogen storage, possibly by sparing glucose from oxidation via preferential consumption of fatty acids.
By examining the metabolic adaptations inspired by AMPK, we see that the overall metabolic flexibility of skeletal muscle is enhanced in a rational and efficient manner. Coordinating increases in beta-oxidation with enhanced respiratory chain capacity ensures that the flow of intermediates from fat oxidation have somewhere to ultimately go; which may be significant, as the incomplete oxidation of fatty acids may have deleterious effects on signaling, and an enhancement of complete fatty acid oxidation is associated with exercise and improved metabolic control [27]. We have not even touched upon the effects that exercise and AMPK have on other tissues, but as one might expect, the consequences are akin to those seen in skeletal muscle, with an enhancement of metabolic flexibility being a key theme.

It bears mentioning that AMPK is also associated with a reduction in protein synthesis – it is, after all, an enzyme induced by fuel depletion, and a lack of fuel is not conducive to anabolic processes. However, resistance training clearly induces hypertrophy in spite of AMPK, so it does seem possible to reap the partitioning benefits of AMPK while stimulating growth. Next month, we will look more closely at the effect of feeding status on AMPK activation, kinases regulating AMPK, and the interaction of diet, exercise type, and AMPK on partitioning.


1. Carlson, C. A. & Kim, K. H. (1973). Regulation of hepatic acetyl coenzyme A carboxylase by phosphorylation and dephosphorylation . J. Biol. Chem, 248:378–80

2. Carling, D., Zammit, V. A. & Hardie, D. G. (1987) A common bicyclic protein kinase cascade inactivates the regulatory enzymes of fatty acid and cholesterol biosynthesis. FEBS Lett , 223(2): 217-22.

3. Hardie, D. G., Carling, D., Sim ATR. (1989) The AMP-activated protein kinase-a multisubstrate regulator of lipid metabolism. Trends Biochem Sci , 14: 20 -23.

4. Hardie, D. G., Carling, D. & Carlson, M. (1998) The AMP-activated/SNF1 protein kinase family: metabolic sensors of the eukaryotic cell? Annu Rev Biochem , 67: 821-55.


5. Hardie, D. G. & Carling, D. (1997) The AMP-activated protein kinase: fuel gauge of the mammalian cell? Eur J Biochem, 246: 259-273.

6. Hardie, D. G. & Hawley, S. A. (2001) AMP-activated protein kinase: the energy charge hypothesis revisited. Bioessays, 23(12): 1112-1119.

7. Vaulont, S., Vasseur-Cognet, M. & Kahn, A. (2000) Glucose regulation of gene transcription. J Biol Chem, 275(41): 31555-8.

8. Kemp, B. E., Mitchelhill, K. I., Stapelton, D., Michell, B. J., Zhi-Ping, C. & Witters, L. A. (1999) Dealing with energy demand: the AMP-activated protein kinase. TIBS, 24: 22-25.

9. Woods, A., Salt, I., Scott, I., Hardie, D. G. & Carling, D. (1996) The á1 and á2 isoforms of the AMP-activated protein kinase have similar activities in rat liver but exhibit differences in substrate specificity in vitro. FEBS Lett, 397: 347-351.

10. Crute, B. E., Seefeld, K., Gamble, J., Kemp, B. E. & Witters, L. A. (1998) Functional domains of the alpha1 catalytic subunit of the AMP-activated protein kinase. J Biol Chem, 273(52): 35347-54.

11. Habinowski, S. A. & Witters, L. A. (2001) The effects of AICAR on adipocytes differentiation of 3T3-L1 cells. Biochem Biophys Res Comm, 286: 852-6.

12. Thorton, C., Snowden, M. A. & Carling, D. (1998) Identification of a novel AMP-activated protein kinase beta subunit isoform that is highly expressed in skeletal muscle. J Biol Chem, 273(20): 12443-12450.

13. Warden, S., Richardson, C., O’Donnell, J. Jr., Stapelton, D., Kemp, B. E. & Witters, L. A. (2001) Post-translational modifications of the â-1 sununit of the AMP-activated protein kinase affect the enzyme activity and cellular localization. Biochem J, 354: 275-283.

14. Mu, J., Brozinick, J. T. Jr., Valladares, O., Bucan, M. & Birnbaum, M. J. (2001) A role for AMP-activated protein kinase in contraction- and hypoxia-regulated glucose transport in skeletal muscle. Mol Cell, 7: 1085-1094.

15. Wojtaszewski, J. F., Birk, J. B., Frosig, C., Holten, M., Pilegaard, H. & Dela. F. (2005) 5’AMP activated protein kinase expression in human skeletal muscle: effects of strength training and type 2 diabetes. J Physiol, 564(Pt 2):563-73

16. Nielsen, J. N., Mustard, K. J., Graham, D. A., Yu, H., MacDonald, C. S., Pilegaard, H., Goodyear, L. J., Hardie, D. G., Richter, E. A. & Wojtaszewski, J. F. (2003) 5′-AMP-activated protein kinase activity and subunit expression in exercise-trained human skeletal muscle. J Appl Physiol, 94(2):631-41.

17. Frosig, C., Jorgensen, S. B., Hardie, D. G., Richter, E. A. & Wojtaszewski, J. F. (2004) 5′-AMP-activated protein kinase activity and protein expression are regulated by endurance training in human skeletal muscle. Am J Physiol Endocrinol Metab, 286(3):E411-7.

18. Clark, S. A., Chen, Z. P., Murphy, K. T., Aughey, R. J., McKenna, M. J/, Kemp, B. E. & Hawley, J. A. (2004) Intensified exercise training does not alter AMPK signaling in human skeletal muscle. Am J Physiol Endocrinol Metab, 286(5):E737-43.

19. Ponticos, M., Lu, Q. L., Morgan, J. E., Hardie, D. G., Partridge, T. A. & Carling, D.
(1998) Dual regulation of the AMP-activated protein kinase provides a novel mechanism for the control of creatine kinase in skeletal muscle. EMBO J, 17(6):1688-99.

20. Chen, Z. P., McConell, G. K., Michell, B. J., Snow, R. J., Canny, B. J. & Kemp, B. E. (2000) AMPK signaling in contracting human skeletal muscle: acetyl-CoA carboxylase and NO synthase phosphorylation. Am J Physiol Endocrinol Metab, 279(5):E1202-6.

21. Fujii, N., Hayashi, T., Hirshman, M. F., Smith, J. T., Habinowski, S. A., Kaijser, L., Mu, J., Ljungqvist, O., Birnbaum, M. J., Witters, L. A., Thorell, A. & Goodyear, L. J. (2000). Exercise induces isoform-specific increase in 5’AMP-activated protein kinase activity in human skeletal muscle. Biochem Biophys Res Commun, 273(3):1150-5.

22. Hardie, D. G. & Pan, D. A. (2002) Regulation of fatty acid synthesis and oxidation by the AMP-activated protein kinase. Biochem Soc Trans, 30(6): 1064-1070.

23. Munday, M. R. (2002) Regulation of mammalian acetyl-CoA carboxylase. Biochem Soc Trans, 30(6): 1059-1064.

24. Page, K. & Lange, Y. (1997) Cell adhesion to fibronectin regulates membrane lipid biosynthesis through 5′-AMP-activated protein kinase. J Biol Chem, 272(31): 19339-42.

25. Zhou, G., Myers, R., Li, Y., Chen, Y., Shen, X., Fenyk-Melody, J., Wu, M., Ventre, J., Doebber, T., Fujii, N., Musi, N., Hirshman, M. F., Goodyear, L. J. & Moller, D. E. (2001) Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest, 108(8): 1167-1174.

26. Lee, W. J., Kim, M., Park, H. S., Kim, H. S., Jeon, M. J., Oh, K. S., Koh, E. H., Won, J. C., Kim, M. S., Oh, G. T., Yoon, M., Lee, K. U. & Park, J. Y. (2006) AMPK activation increases fatty acid oxidation in skeletal muscle by activating PPARalpha and PGC-1.Biochem Biophys Res Commun, 340(1):291-5.

27. Koves, T. R., Li, P., An, J., Akimoto, T., Slentz, D., Ilkayeva, O., Dohm, G. L., Yan, Z., Newgard, C. B. & Muoio, D. M. (2005) Peroxisome proliferator-activated receptor-gamma co-activator 1alpha-mediated metabolic remodeling of skeletal myocytes mimics exercise training and reverses lipid-induced mitochondrial inefficiency. J Biol Chem, 280(39):33588-98.

28. Adams, J., Chen, Z. P., Van Denderen, B. J., Morton, C. J., Parker, M. W., Witters, L. A., Stapleton, D. & Kemp, B. E. (2004) Intrasteric control of AMPK via the gamma1 subunit AMP allosteric regulatory site. Protein Sci, 13(1):155-65.

29. Eijnde, B. O., Derave, W., Wojtaszewski, J. F., Richter, E. A. & Hespel, P. (2005) AMP kinase expression and activity in human skeletal muscle: effects of immobilization, retraining, and creatine supplementation. J Appl Physiol, 98(4):1228-33

PCT + AI Stack + 2 items
someone from Concord
Total order for 54.45 USD
someone from Waco
Total order for 89.45 USD
Rad Bod Stack + 5 items
someone from Killeen
Total order for 134.90 USD
someone from Lees Summit
Total order for 64.49 USD
Liquid Labs T2
someone from Elnhurst
Total order for 72.97 USD